‘segmentation’

Comparison of existing and proposed labeling approaches. In a traditional context, either an expert rater with extensive anatomical knowledge evaluates each dataset (A) or a small set of well-training domain experts who have been instructed by an anatomical expert (B) label each image. Intra- and inter-rater reproducibility analyses are typically performed a per-protocol basis rather than on all datasets. In the proposed WebMILL approach (C), a computer system divides the set of images to be labeled into simple puzzles consisting of a piece of a larger volume and distributes these challenges to a distributed collection of minimally training individuals. Each piece is labeled multiple times by multiple raters. A statistical fusion process simultaneously estimates the true label for each pixel and performance characteristics of each rater.

Foibles, Follies, and Fusion: Web-Based Collaboration for Medical Image Labeling

Jan. 31, 2012—Bennett A Landman, Andrew J Asman, Andrew G Scoggins, John A Bogovic, Joshua A Stein; Jerry L Prince, “Foibles, Follies, and Fusion: Web-Based Collaboration for Medical Image Labeling”, NeuroImage. 2012 Jan 2;59(1):530-9. PMC3195954 † Full text: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3195954/ Abstract Labels that identify specific anatomical and functional structures within medical images are essential to the characterization of...

Statistical Fusion of Surface Labels provided by Multiple Raters

Mar. 1, 2010—Bogovic, B. A. Landman, P.-L. Bazin, and J. L. Prince. “Statistical Fusion of Surface Labels provided by Multiple Raters, Over-complete, and Ancillary Data”, In Proceedings of the SPIE Medical Imaging Conference. San Diego, CA, February 2010 PMC2997739 Full Text: https://www.ncbi.nlm.nih.gov/pubmed/?term=Statistical+Fusion+of+Surface+Labels+provided+by+Multiple+Raters Abstract Studies of the size and morphology of anatomical structures rely on accurate and reproducible delineation...

Simulations with traditional random raters. Coronal sections of the three-dimensional volume show the high resolution MRI image (A), manually drawn truth model (B), an example delineation from one random traditional rater (C), and the results of a STAPLE recombination of three label sets (D). STAPLER enables fusion of label sets when raters provide only partial datasets, but performance suffers with decreasing overlap (E). With training data (F), STAPLER improved the performance even with each rater labeling only a small portion of the dataset. Box plots in E and F show mean, quartiles, range up to 1.5σ, and outliers. The highlighted plot in E indicates the simulation for which STAPLER was equivalent to STAPLE--i.e., all raters provide a complete set of labels.

Simultaneous Truth and Performance Level Estimation with Incomplete, Over-complete, and Ancillary Data

Mar. 1, 2010—A. Landman, H. Wan, J. Bogovic, and J. L. Prince. “Simultaneous Truth and Performance Level Estimation with Incomplete, Over-complete, and Ancillary Data”, In Proceedings of the SPIE Medical Imaging Conference. San Diego, CA, February 2010 (Oral Presentation) PMC2917119 Full Text:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2917119/ Abstract Image labeling and parcellation are critical tasks for the assessment of volumetric and morphometric...